In [1]:
#主要是用于数据处理、可视化、建模、评分等
import pandas as pd
import numpy as np
pd.set_option('display.max_columns', 32)

# 可视化
import matplotlib.pyplot as plt
import seaborn as sns
plt.style.use('fivethirtyeight')
import plotly.express as px
%matplotlib inline

# 缺失值可视化
import missingno as msno
# 地图可视化
import folium
from folium.plugins import HeatMap

# 建模相关
from sklearn.model_selection import train_test_split, GridSearchCV
from sklearn.preprocessing import StandardScaler
from sklearn import metrics
from sklearn.metrics import classification_report, confusion_matrix, ConfusionMatrixDisplay
from sklearn.metrics import accuracy_score, recall_score, roc_auc_score, precision_score, f1_score
from sklearn.metrics import accuracy_score, confusion_matrix, classification_report
from sklearn.linear_model import LogisticRegression
from sklearn.neighbors import KNeighborsClassifier
from sklearn.svm import SVC
from sklearn.tree import DecisionTreeClassifier
from sklearn.ensemble import RandomForestClassifier
from sklearn.ensemble import RandomForestClassifier
from sklearn.ensemble import AdaBoostClassifier
from sklearn.ensemble import GradientBoostingClassifier
from xgboost import XGBClassifier
from catboost import CatBoostClassifier
from sklearn.ensemble import ExtraTreesClassifier
from lightgbm import LGBMClassifier
from sklearn.ensemble import VotingClassifier

# 忽略警告
import warnings
warnings.filterwarnings('ignore')
In [2]:
df=pd.read_csv('hotel_bookings.csv')
df.head()
Out[2]:
hotel is_canceled lead_time arrival_date_year arrival_date_month arrival_date_week_number arrival_date_day_of_month stays_in_weekend_nights stays_in_week_nights adults children babies meal country market_segment distribution_channel is_repeated_guest previous_cancellations previous_bookings_not_canceled reserved_room_type assigned_room_type booking_changes deposit_type agent company days_in_waiting_list customer_type adr required_car_parking_spaces total_of_special_requests reservation_status reservation_status_date
0 Resort Hotel 0 342 2015 July 27 1 0 0 2 0.0 0 BB PRT Direct Direct 0 0 0 C C 3 No Deposit NaN NaN 0 Transient 0.0 0 0 Check-Out 2015-07-01
1 Resort Hotel 0 737 2015 July 27 1 0 0 2 0.0 0 BB PRT Direct Direct 0 0 0 C C 4 No Deposit NaN NaN 0 Transient 0.0 0 0 Check-Out 2015-07-01
2 Resort Hotel 0 7 2015 July 27 1 0 1 1 0.0 0 BB GBR Direct Direct 0 0 0 A C 0 No Deposit NaN NaN 0 Transient 75.0 0 0 Check-Out 2015-07-02
3 Resort Hotel 0 13 2015 July 27 1 0 1 1 0.0 0 BB GBR Corporate Corporate 0 0 0 A A 0 No Deposit 304.0 NaN 0 Transient 75.0 0 0 Check-Out 2015-07-02
4 Resort Hotel 0 14 2015 July 27 1 0 2 2 0.0 0 BB GBR Online TA TA/TO 0 0 0 A A 0 No Deposit 240.0 NaN 0 Transient 98.0 0 1 Check-Out 2015-07-03
In [3]:
# 1、字段信息

df.columns
Out[3]:
Index(['hotel', 'is_canceled', 'lead_time', 'arrival_date_year',
       'arrival_date_month', 'arrival_date_week_number',
       'arrival_date_day_of_month', 'stays_in_weekend_nights',
       'stays_in_week_nights', 'adults', 'children', 'babies', 'meal',
       'country', 'market_segment', 'distribution_channel',
       'is_repeated_guest', 'previous_cancellations',
       'previous_bookings_not_canceled', 'reserved_room_type',
       'assigned_room_type', 'booking_changes', 'deposit_type', 'agent',
       'company', 'days_in_waiting_list', 'customer_type', 'adr',
       'required_car_parking_spaces', 'total_of_special_requests',
       'reservation_status', 'reservation_status_date'],
      dtype='object')

字段的具体中文含义：

hotel 酒店

is_canceled 是否取消

lead_time 预订时间

arrival_date_year 入住年份

arrival_date_month 入住月份

arrival_date_week_number 入住周次

arrival_date_day_of_month 入住天号

stays_in_weekend_nights 周末夜晚数

stays_in_week_nights 工作日夜晚数 adults 成人数量

children 儿童数量

babies 幼儿数量

meal 餐食

country 国家

market_segment 细分市场

distribution_channel 分销渠道

is_repeated_guest 是否是回头客

previous_cancellations 先前取消数

previous_bookings_not_canceled 先前未取消数

reserved_room_type 预订房间类型

assigned_room_type 实际房间类型

booking_changes 预订更改数

deposit_type 押金方式

agent 代理

company 公司

days_in_waiting_list 排队天数

customer_type 客户类型

adr 每日房间均价 （Average Daily Rate）

required_car_parking_spaces 停车位数量

total_of_special_requests 特殊需求数(例如高层或双床)

reservation_status 订单状态

reservation_status_date 订单状态确定日期

In [4]:
# 2、总字段个数
len(df.columns)
Out[4]:
32
In [5]:
# 3、字段类型
df.dtypes
Out[5]:
hotel                              object
is_canceled                         int64
lead_time                           int64
arrival_date_year                   int64
arrival_date_month                 object
arrival_date_week_number            int64
arrival_date_day_of_month           int64
stays_in_weekend_nights             int64
stays_in_week_nights                int64
adults                              int64
children                          float64
babies                              int64
meal                               object
country                            object
market_segment                     object
distribution_channel               object
is_repeated_guest                   int64
previous_cancellations              int64
previous_bookings_not_canceled      int64
reserved_room_type                 object
assigned_room_type                 object
booking_changes                     int64
deposit_type                       object
agent                             float64
company                           float64
days_in_waiting_list                int64
customer_type                      object
adr                               float64
required_car_parking_spaces         int64
total_of_special_requests           int64
reservation_status                 object
reservation_status_date            object
dtype: object
In [6]:
# 4、不同类型字段统计
df.dtypes.value_counts()
Out[6]:
int64      16
object     12
float64     4
dtype: int64
In [7]:
# 5、数据量
df.shape
Out[7]:
(119390, 32)
In [8]:
# 6、描述统计信息
df.describe()
Out[8]:
is_canceled lead_time arrival_date_year arrival_date_week_number arrival_date_day_of_month stays_in_weekend_nights stays_in_week_nights adults children babies is_repeated_guest previous_cancellations previous_bookings_not_canceled booking_changes agent company days_in_waiting_list adr required_car_parking_spaces total_of_special_requests
count 119390.000000 119390.000000 119390.000000 119390.000000 119390.000000 119390.000000 119390.000000 119390.000000 119386.000000 119390.000000 119390.000000 119390.000000 119390.000000 119390.000000 103050.000000 6797.000000 119390.000000 119390.000000 119390.000000 119390.000000
mean 0.370416 104.011416 2016.156554 27.165173 15.798241 0.927599 2.500302 1.856403 0.103890 0.007949 0.031912 0.087118 0.137097 0.221124 86.693382 189.266735 2.321149 101.831122 0.062518 0.571363
std 0.482918 106.863097 0.707476 13.605138 8.780829 0.998613 1.908286 0.579261 0.398561 0.097436 0.175767 0.844336 1.497437 0.652306 110.774548 131.655015 17.594721 50.535790 0.245291 0.792798
min 0.000000 0.000000 2015.000000 1.000000 1.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 1.000000 6.000000 0.000000 -6.380000 0.000000 0.000000
25% 0.000000 18.000000 2016.000000 16.000000 8.000000 0.000000 1.000000 2.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 9.000000 62.000000 0.000000 69.290000 0.000000 0.000000
50% 0.000000 69.000000 2016.000000 28.000000 16.000000 1.000000 2.000000 2.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 14.000000 179.000000 0.000000 94.575000 0.000000 0.000000
75% 1.000000 160.000000 2017.000000 38.000000 23.000000 2.000000 3.000000 2.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 229.000000 270.000000 0.000000 126.000000 0.000000 1.000000
max 1.000000 737.000000 2017.000000 53.000000 31.000000 19.000000 50.000000 55.000000 10.000000 10.000000 1.000000 26.000000 72.000000 21.000000 535.000000 543.000000 391.000000 5400.000000 8.000000 5.000000
In [9]:
# 7、数据完整信息
df.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 119390 entries, 0 to 119389
Data columns (total 32 columns):
 #   Column                          Non-Null Count   Dtype  
---  ------                          --------------   -----  
 0   hotel                           119390 non-null  object 
 1   is_canceled                     119390 non-null  int64  
 2   lead_time                       119390 non-null  int64  
 3   arrival_date_year               119390 non-null  int64  
 4   arrival_date_month              119390 non-null  object 
 5   arrival_date_week_number        119390 non-null  int64  
 6   arrival_date_day_of_month       119390 non-null  int64  
 7   stays_in_weekend_nights         119390 non-null  int64  
 8   stays_in_week_nights            119390 non-null  int64  
 9   adults                          119390 non-null  int64  
 10  children                        119386 non-null  float64
 11  babies                          119390 non-null  int64  
 12  meal                            119390 non-null  object 
 13  country                         118902 non-null  object 
 14  market_segment                  119390 non-null  object 
 15  distribution_channel            119390 non-null  object 
 16  is_repeated_guest               119390 non-null  int64  
 17  previous_cancellations          119390 non-null  int64  
 18  previous_bookings_not_canceled  119390 non-null  int64  
 19  reserved_room_type              119390 non-null  object 
 20  assigned_room_type              119390 non-null  object 
 21  booking_changes                 119390 non-null  int64  
 22  deposit_type                    119390 non-null  object 
 23  agent                           103050 non-null  float64
 24  company                         6797 non-null    float64
 25  days_in_waiting_list            119390 non-null  int64  
 26  customer_type                   119390 non-null  object 
 27  adr                             119390 non-null  float64
 28  required_car_parking_spaces     119390 non-null  int64  
 29  total_of_special_requests       119390 non-null  int64  
 30  reservation_status              119390 non-null  object 
 31  reservation_status_date         119390 non-null  object 
dtypes: float64(4), int64(16), object(12)
memory usage: 29.1+ MB

缺失值信息¶

统计每个字段缺失值信息¶

统计每个字段的缺失值数量及比例

In [10]:
null_df = pd.DataFrame({"Null Values": df.isnull().sum(),
                         "Percentage Null Values": (df.isnull().sum()) / (df.shape[0]) * 100
                         })

null_df
Out[10]:
Null Values Percentage Null Values
hotel 0 0.000000
is_canceled 0 0.000000
lead_time 0 0.000000
arrival_date_year 0 0.000000
arrival_date_month 0 0.000000
arrival_date_week_number 0 0.000000
arrival_date_day_of_month 0 0.000000
stays_in_weekend_nights 0 0.000000
stays_in_week_nights 0 0.000000
adults 0 0.000000
children 4 0.003350
babies 0 0.000000
meal 0 0.000000
country 488 0.408744
market_segment 0 0.000000
distribution_channel 0 0.000000
is_repeated_guest 0 0.000000
previous_cancellations 0 0.000000
previous_bookings_not_canceled 0 0.000000
reserved_room_type 0 0.000000
assigned_room_type 0 0.000000
booking_changes 0 0.000000
deposit_type 0 0.000000
agent 16340 13.686238
company 112593 94.306893
days_in_waiting_list 0 0.000000
customer_type 0 0.000000
adr 0 0.000000
required_car_parking_spaces 0 0.000000
total_of_special_requests 0 0.000000
reservation_status 0 0.000000
reservation_status_date 0 0.000000

缺失值可视化¶

将缺失值信息进行可视化展示：¶

In [11]:
msno.bar(df, color="blue")

plt.show()

缺失值处理

1、字段 children和字段country 缺失值比例都不到1%，比例很小；我们直接把缺失值的部分删除

In [12]:
# 把非缺失值的数据筛选出来

df = df[df["country"].isnull() == False]
df = df[df["children"].isnull() == False]

df.head()
Out[12]:
hotel is_canceled lead_time arrival_date_year arrival_date_month arrival_date_week_number arrival_date_day_of_month stays_in_weekend_nights stays_in_week_nights adults children babies meal country market_segment distribution_channel is_repeated_guest previous_cancellations previous_bookings_not_canceled reserved_room_type assigned_room_type booking_changes deposit_type agent company days_in_waiting_list customer_type adr required_car_parking_spaces total_of_special_requests reservation_status reservation_status_date
0 Resort Hotel 0 342 2015 July 27 1 0 0 2 0.0 0 BB PRT Direct Direct 0 0 0 C C 3 No Deposit NaN NaN 0 Transient 0.0 0 0 Check-Out 2015-07-01
1 Resort Hotel 0 737 2015 July 27 1 0 0 2 0.0 0 BB PRT Direct Direct 0 0 0 C C 4 No Deposit NaN NaN 0 Transient 0.0 0 0 Check-Out 2015-07-01
2 Resort Hotel 0 7 2015 July 27 1 0 1 1 0.0 0 BB GBR Direct Direct 0 0 0 A C 0 No Deposit NaN NaN 0 Transient 75.0 0 0 Check-Out 2015-07-02
3 Resort Hotel 0 13 2015 July 27 1 0 1 1 0.0 0 BB GBR Corporate Corporate 0 0 0 A A 0 No Deposit 304.0 NaN 0 Transient 75.0 0 0 Check-Out 2015-07-02
4 Resort Hotel 0 14 2015 July 27 1 0 2 2 0.0 0 BB GBR Online TA TA/TO 0 0 0 A A 0 No Deposit 240.0 NaN 0 Transient 98.0 0 1 Check-Out 2015-07-03

2、字段company缺失值比例高达94.3%，我们考虑直接删除该字段：

In [13]:
df.drop("company", axis=1, inplace=True)

3、字段agent(代理商费用)的缺失值为13.68%，处理为：

In [14]:
# 1、先查看字段具体信息
df["agent"].value_counts()
Out[14]:
9.0      31959
240.0    13871
1.0       7191
14.0      3638
7.0       3539
         ...  
70.0         1
93.0         1
54.0         1
497.0        1
59.0         1
Name: agent, Length: 332, dtype: int64

我们可以考虑使用的值来进行填充，比如：

0：无法确定缺失值的具体数据

9：众数

均值：字段现有值的均值

在这里我们考虑使用0来进行填充：

In [15]:
df["agent"].fillna(0,inplace=True)

特殊处理¶

处理1：入住人数不能为0

考虑到一个房间中adults、children和babies的数量不能同时为0：|

In [16]:
special = (df["children"] == 0) & (df.adults == 0) & (df.babies == 0)
special.head()
Out[16]:
0    False
1    False
2    False
3    False
4    False
dtype: bool
In [17]:
# 排除特殊情况
df = df[~special
       3
       ]

处理2：adr（日均价）¶

取值不能为负数

最大值为5400，可以判断属于异常值

In [18]:
df["adr"].value_counts().sort_index()
Out[18]:
-6.38          1
 0.00       1799
 0.26          1
 0.50          1
 1.00         14
            ... 
 450.00        1
 451.50        1
 508.00        1
 510.00        1
 5400.00       1
Name: adr, Length: 8857, dtype: int64
In [19]:
#通过小提琴图来查看数据的分布情况：处理前明显有离群点
px.violin(y=df["adr"])   # 处理前
In [20]:
px.box(df,y="adr")
In [21]:
# 删除大于1000的信息  df = df.drop(df[df.adr >1000].index)

df = df[(df["adr"] >= 0) & (df["adr"] < 5400)]  # 排除异常值
In [22]:
px.violin(y=df["adr"])  # 删除后
In [23]:
px.box(df,y="adr",color="hotel")   # 删除后

数据EDA-Exploratory Data Analysis¶

取消和未取消的顾客数对比¶

In [24]:
df["is_canceled"].value_counts()
Out[24]:
0    74589
1    44137
Name: is_canceled, dtype: int64
In [25]:
# 取消和未取消人数对比  0-未取消 1-取消
sns.countplot(df["is_canceled"])

plt.show()
In [26]:
#未取消的顾客来自哪里？
data = df[df.is_canceled == 0]  # 未取消的数据
In [27]:
number_no_canceled = data["country"].value_counts().reset_index()
number_no_canceled.columns = ["country", "number_of_no_canceled"]

number_no_canceled
Out[27]:
country number_of_no_canceled
0 PRT 20977
1 GBR 9667
2 FRA 8468
3 ESP 6383
4 DEU 6067
... ... ...
160 BHR 1
161 DJI 1
162 MLI 1
163 NPL 1
164 FRO 1

165 rows × 2 columns

In [28]:
# 地图可视化

basemap = folium.Map()
guests_map = px.choropleth(number_no_canceled, # 传入数据
                           locations = number_no_canceled['country'],  # 地理位置
                           color = number_no_canceled['number_of_no_canceled'],  # 颜色取值
                           hover_name = number_no_canceled['country'])  # 悬停信息
guests_map.show()

结论1：预订的顾客主要是来自Portugal，大部分是欧洲的国家¶

In [29]:
#房间的每日均价是多少？

px.box(data,  # 数据
       x="reserved_room_type",  # x
       y="adr", # y
       color="hotel",  # 颜色
       template="plotly_dark",  # 主题
       category_orders={"reserved_room_type":["A","B","C","D","E","F","G","H","L"]} # 指定排列顺序
      )

结论2：每个房间的均价还是取决于它的类型和标准差¶

全年每晚的价格是多少？

两种不同类型酒店的全年均价变化

In [30]:
data_resort = data[data["hotel"] == "Resort Hotel"]
data_city = data[data["hotel"] == "City Hotel"]
In [31]:
resort_hotel = data_resort.groupby(['arrival_date_month'])['adr'].mean().reset_index()
city_hotel = data_city.groupby(['arrival_date_month'])['adr'].mean().reset_index()
city_hotel
Out[31]:
arrival_date_month adr
0 April 111.962267
1 August 118.674598
2 December 88.401855
3 February 86.520062
4 January 82.330983
5 July 115.818019
6 June 117.874360
7 March 90.658533
8 May 120.669827
9 November 86.946178
10 October 102.004672
11 September 112.776582
In [32]:
# 合并两个数据

total_hotel = pd.merge(resort_hotel, city_hotel,
                        on="arrival_date_month"
                        )
total_hotel.columns = ["month","price_resort","price_city"]
total_hotel
Out[32]:
month price_resort price_city
0 April 76.176140 111.962267
1 August 181.348715 118.674598
2 December 68.634584 88.401855
3 February 54.468180 86.520062
4 January 48.994018 82.330983
5 July 150.483553 115.818019
6 June 108.133154 117.874360
7 March 57.433661 90.658533
8 May 76.734804 120.669827
9 November 48.882907 86.946178
10 October 61.932777 102.004672
11 September 96.288897 112.776582

为了让月份按照正常时间排序，安装两个包：

jupyter notebook直接安装：前面要加!

In [33]:
#!pip install sort-dataframeby-monthorweek
#!pip install sorted-months-weekdays
In [34]:
import sort_dataframeby_monthorweek as sd

#  自定义排序函数
def sort_month(df, column):
    result = sd.Sort_Dataframeby_Month(df,column)
    return result
In [35]:
new_total_hotel = sort_month(total_hotel, "month")
new_total_hotel
Out[35]:
month price_resort price_city
0 January 48.994018 82.330983
1 February 54.468180 86.520062
2 March 57.433661 90.658533
3 April 76.176140 111.962267
4 May 76.734804 120.669827
5 June 108.133154 117.874360
6 July 150.483553 115.818019
7 August 181.348715 118.674598
8 September 96.288897 112.776582
9 October 61.932777 102.004672
10 November 48.882907 86.946178
11 December 68.634584 88.401855
In [36]:
fig = px.line(new_total_hotel,
        x = "month",
        y = ["price_resort", "price_city"],
        title = "Price of per night over the Months",
        template = "plotly_dark"      
       )

fig.show()

结论：¶

Resort Hotel在夏季的价格明显比 City Hotel的价格高

City Hotel的价格变化相对更小。但是City Hotel的价格从4月开始就已经很高，一直持续到9月份

KDE图¶

KDE(Kernel Density Estimation，核密度图)，可以认为是对直方图的加窗平滑。通过KDE分布图场内看数据在不同情形下的分布

In [37]:
plt.figure(figsize=(6,3), dpi=150)

ax = sns.kdeplot(new_total_hotel["price_resort"], 
                 color="green", 
                 shade=True)

ax = sns.kdeplot(new_total_hotel["price_city"], 
                 color="blue", 
                 shade=True)

ax.set_xlabel("month")
ax.set_ylabel("Price per night over the month")
ax = ax.legend(["Resort","City"])

最为繁忙的季节-the most busy months¶

In [38]:
resort_guests = data_resort['arrival_date_month'].value_counts().reset_index()
resort_guests.columns=['Month','No_Resort_Guests']

city_guests = data_city['arrival_date_month'].value_counts().reset_index()
city_guests.columns=['Month','No_City_Guests']

# 合并两份DataFrame
final_guests = pd.merge(resort_guests, city_guests)
In [39]:
#同样的将月份进行排序处理
new_final_guests = sort_month(final_guests, "Month")
new_final_guests
Out[39]:
Month No_Resort_Guests No_City_Guests
0 January 1812 2249
1 February 2253 3051
2 March 2516 4049
3 April 2518 4010
4 May 2523 4568
5 June 2026 4358
6 July 3110 4770
7 August 3237 5367
8 September 2077 4283
9 October 2528 4326
10 November 1937 2674
11 December 1970 2377
In [40]:
fig = px.line(new_final_guests,
        x = "Month",
        y = ["No_Resort_Guests", "No_City_Guests"],
        title = "No of per Month",
        template = "plotly_dark"      
       )

fig.show()

结论：¶

很明显：City Hotel的人数是高于Resort Hotel，更受欢迎

City Hotel在7-8月份的时候，尽管价格高（上图），但人数也达到了峰值

两个Hotel在冬季的顾客都是很少的

顾客停留多久？¶

In [41]:
data["total_nights"] = data['stays_in_weekend_nights'] + data['stays_in_week_nights']
data.head()
Out[41]:
hotel is_canceled lead_time arrival_date_year arrival_date_month arrival_date_week_number arrival_date_day_of_month stays_in_weekend_nights stays_in_week_nights adults children babies meal country market_segment distribution_channel is_repeated_guest previous_cancellations previous_bookings_not_canceled reserved_room_type assigned_room_type booking_changes deposit_type agent days_in_waiting_list customer_type adr required_car_parking_spaces total_of_special_requests reservation_status reservation_status_date total_nights
0 Resort Hotel 0 342 2015 July 27 1 0 0 2 0.0 0 BB PRT Direct Direct 0 0 0 C C 3 No Deposit 0.0 0 Transient 0.0 0 0 Check-Out 2015-07-01 0
1 Resort Hotel 0 737 2015 July 27 1 0 0 2 0.0 0 BB PRT Direct Direct 0 0 0 C C 4 No Deposit 0.0 0 Transient 0.0 0 0 Check-Out 2015-07-01 0
2 Resort Hotel 0 7 2015 July 27 1 0 1 1 0.0 0 BB GBR Direct Direct 0 0 0 A C 0 No Deposit 0.0 0 Transient 75.0 0 0 Check-Out 2015-07-02 1
3 Resort Hotel 0 13 2015 July 27 1 0 1 1 0.0 0 BB GBR Corporate Corporate 0 0 0 A A 0 No Deposit 304.0 0 Transient 75.0 0 0 Check-Out 2015-07-02 1
4 Resort Hotel 0 14 2015 July 27 1 0 2 2 0.0 0 BB GBR Online TA TA/TO 0 0 0 A A 0 No Deposit 240.0 0 Transient 98.0 0 1 Check-Out 2015-07-03 2
In [42]:
#两个不同酒店在不同停留时间下的统计：

stay_groupby = (data.groupby(['total_nights', 'hotel'])["is_canceled"]
                .agg("count")
                .reset_index()
                .rename(columns={"is_canceled":"Number of stays"}))

stay_groupby.head()
Out[42]:
total_nights hotel Number of stays
0 0 City Hotel 251
1 0 Resort Hotel 366
2 1 City Hotel 9155
3 1 Resort Hotel 6368
4 2 City Hotel 10983
In [43]:
stay_groupby = (data.groupby(['total_nights', 'hotel'])["is_canceled"]
                .agg("count")
                .reset_index()
                .rename(columns={"is_canceled":"Number of stays"}))

stay_groupby.head()
Out[43]:
total_nights hotel Number of stays
0 0 City Hotel 251
1 0 Resort Hotel 366
2 1 City Hotel 9155
3 1 Resort Hotel 6368
4 2 City Hotel 10983
In [44]:
fig = px.bar(stay_groupby,
       x = "total_nights",
       y = "Number of stays",
       color = "hotel",
       barmode = "group"
      )

fig.show()

数据预处理-Data Pre Processing¶

相关性判断¶

In [45]:
plt.figure (figsize=(24,12))

corr = df.corr()
sns.heatmap(corr, annot = True, linewidths = 1)
plt.show()
In [46]:
#查看每个特征和目标变量is_canceled的相关系数的绝对值，并降序排列：
corr_with_iscanceled = df.corr()["is_canceled"].abs().sort_values(ascending=False)

corr_with_iscanceled
Out[46]:
is_canceled                       1.000000
lead_time                         0.291619
total_of_special_requests         0.235923
required_car_parking_spaces       0.195013
booking_changes                   0.145139
previous_cancellations            0.109911
is_repeated_guest                 0.084115
adults                            0.056129
previous_bookings_not_canceled    0.055494
days_in_waiting_list              0.054114
agent                             0.046842
adr                               0.045910
babies                            0.032605
stays_in_week_nights              0.024825
arrival_date_year                 0.016419
arrival_date_week_number          0.007668
arrival_date_day_of_month         0.006022
children                          0.004536
stays_in_weekend_nights           0.002192
Name: is_canceled, dtype: float64
In [47]:
#删除无效字段
no_use_col = ['arrival_date_year', 'assigned_room_type',
             'booking_changes','reservation_status', 
             'country', 'days_in_waiting_list']
In [48]:
df.drop(no_use_col, axis=1, inplace=True)

特征工程¶

离散型变量处理

In [49]:
df["hotel"].dtype # Series型数据的字段类型
Out[49]:
dtype('O')
In [50]:
cat_cols = [col for col in df.columns if df[col].dtype == "O"]
cat_cols
Out[50]:
['hotel',
 'arrival_date_month',
 'meal',
 'market_segment',
 'distribution_channel',
 'reserved_room_type',
 'deposit_type',
 'customer_type',
 'reservation_status_date']
In [51]:
cat_df = df[cat_cols]
In [52]:
cat_df.dtypes
Out[52]:
hotel                      object
arrival_date_month         object
meal                       object
market_segment             object
distribution_channel       object
reserved_room_type         object
deposit_type               object
customer_type              object
reservation_status_date    object
dtype: object
In [53]:
# 1、转成时间类型数据

cat_df['reservation_status_date'] = pd.to_datetime(cat_df['reservation_status_date'])
In [54]:
#  2、提取年月日

cat_df["year"] = cat_df['reservation_status_date'].dt.year
cat_df['month'] = cat_df['reservation_status_date'].dt.month
cat_df['day'] = cat_df['reservation_status_date'].dt.day
In [55]:
# 3、删除无效字段
cat_df.drop(['reservation_status_date','arrival_date_month'], axis=1, inplace=True)
In [56]:
# 4、每个字段的唯一值

for col in cat_df.columns:
    print(f"{col}: \n{cat_df[col].unique()}\n")

hotel: 
['Resort Hotel' 'City Hotel']

meal: 
['BB' 'FB' 'HB' 'SC' 'Undefined']

market_segment: 
['Direct' 'Corporate' 'Online TA' 'Offline TA/TO' 'Complementary' 'Groups'
 'Aviation']

distribution_channel: 
['Direct' 'Corporate' 'TA/TO' 'Undefined' 'GDS']

reserved_room_type: 
['C' 'A' 'D' 'E' 'G' 'F' 'H' 'L' 'B']

deposit_type: 
['No Deposit' 'Refundable' 'Non Refund']

customer_type: 
['Transient' 'Contract' 'Transient-Party' 'Group']

year: 
[2015 2014 2016 2017]

month: 
[ 7  5  4  6  3  8  9  1 11 10 12  2]

day: 
[ 1  2  3  6 22 23  5  7  8 11 16 29 19 18  9 13  4 12 26 17 15 10 20 14
 30 28 25 21 27 24 31]

特征编码¶

In [57]:
# 酒店
cat_df['hotel'] = cat_df['hotel'].map({'Resort Hotel' : 0, 
                                       'City Hotel' : 1})
# 餐食
cat_df['meal'] = cat_df['meal'].map({'BB' : 0, 
                                     'FB': 1, 
                                     'HB': 2, 
                                     'SC': 3, 
                                     'Undefined': 4})
# 细分市场
cat_df['market_segment'] = (cat_df['market_segment']
                            .map({'Direct': 0,
                                 'Corporate':1,
                                 'Online TA':2, 
                                 'Offline TA/TO': 3,
                                 'Complementary': 4,
                                 'Groups': 5,
                                 'Undefined': 6,
                                 'Aviation': 7}))
# 分销渠道
cat_df['distribution_channel'] = (cat_df['distribution_channel']
                                  .map({'Direct': 0,
                                        'Corporate': 1,
                                        'TA/TO': 2, 
                                        'Undefined': 3,
                                        'GDS': 4}))
# 预订房间类型
cat_df['reserved_room_type'] = (cat_df['reserved_room_type']
                                .map({'C': 0, 
                                      'A': 1, 
                                      'D': 2, 
                                      'E': 3, 
                                      'G': 4, 
                                      'F': 5, 
                                      'H': 6,
                                      'L': 7, 
                                      'B': 8}))
# 押金方式
cat_df['deposit_type'] = (cat_df['deposit_type']
                          .map({'No Deposit': 0, 
                                'Refundable': 1, 
                                'Non Refund': 3}))
# 顾客类型
cat_df['customer_type'] = (cat_df['customer_type']
                           .map({'Transient': 0, 
                                 'Contract': 1, 
                                 'Transient-Party': 2, 
                                 'Group': 3})
                          )
# 年份
cat_df['year'] = cat_df['year'].map({2015: 0, 2014: 1, 2016: 2, 2017: 3})

连续型变量处理¶

In [58]:
num_df = df.drop(columns=cat_cols,axis=1)

num_df.drop("is_canceled",axis=1,inplace=True)
In [59]:
# 方差偏大的字段进行对数化处理
log_col = ["lead_time","arrival_date_week_number","arrival_date_day_of_month","agent","adr"]

for col in log_col:
    num_df[col] = np.log(num_df[col] + 1)
    
num_df.head()
Out[59]:
lead_time arrival_date_week_number arrival_date_day_of_month stays_in_weekend_nights stays_in_week_nights adults children babies is_repeated_guest previous_cancellations previous_bookings_not_canceled agent adr required_car_parking_spaces total_of_special_requests
0 5.837730 3.332205 0.693147 0 0 2 0.0 0 0 0 0 0.000000 0.000000 0 0
1 6.603944 3.332205 0.693147 0 0 2 0.0 0 0 0 0 0.000000 0.000000 0 0
2 2.079442 3.332205 0.693147 0 1 1 0.0 0 0 0 0 0.000000 4.330733 0 0
3 2.639057 3.332205 0.693147 0 1 1 0.0 0 0 0 0 5.720312 4.330733 0 0
4 2.708050 3.332205 0.693147 0 2 2 0.0 0 0 0 0 5.484797 4.595120 0 1

建模¶

合并两份df¶

In [60]:
X = pd.concat([cat_df, num_df], axis=1)
y = df["is_canceled"]
In [61]:
print(X.shape)
print(y.shape)

(118726, 25)
(118726,)
In [62]:
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.2,random_state=412)

建模1：逻辑回归¶

In [63]:
# 模型实例化
lr = LogisticRegression()
lr.fit(X_train, y_train)

# 预测值
y_pred_lr = lr.predict(X_test)

# 分类问题不同评价指标
acc_lr = accuracy_score(y_test, y_pred_lr)
conf = confusion_matrix(y_test, y_pred_lr)
clf_report = classification_report(y_test, y_pred_lr)

print(f"Accuracy Score of Logistic Regression is : {acc_lr}")
print(f"Confusion Matrix : \n{conf}")
print(f"Classification Report : \n{clf_report}")

Accuracy Score of Logistic Regression is : 0.8134843763160111
Confusion Matrix : 
[[14172   797]
 [ 3632  5145]]
Classification Report : 
              precision    recall  f1-score   support

           0       0.80      0.95      0.86     14969
           1       0.87      0.59      0.70      8777

    accuracy                           0.81     23746
   macro avg       0.83      0.77      0.78     23746
weighted avg       0.82      0.81      0.80     23746
In [64]:
# 混淆矩阵可视化

classes = ["0","1"]

disp = ConfusionMatrixDisplay(confusion_matrix=conf, display_labels=classes)
disp.plot(
    include_values=True,            # 混淆矩阵每个单元格上显示具体数值
    cmap="GnBu",                 # matplotlib识别的颜色图
    ax=None,
    xticks_rotation="horizontal",
    values_format="d"
)

plt.show()
In [65]:
#KNN 模型
knn = KNeighborsClassifier()
knn.fit(X_train, y_train)

y_pred= knn.predict(X_test)

acc_knn = accuracy_score(y_test, y_pred)
conf = confusion_matrix(y_test, y_pred)
clf_report = classification_report(y_test, y_pred)
In [66]:
#决策树模型
dtc = DecisionTreeClassifier()
dtc.fit(X_train, y_train)

y_pred_dtc = dtc.predict(X_test)

acc_dtc = accuracy_score(y_test, y_pred_dtc)
conf = confusion_matrix(y_test, y_pred_dtc)
clf_report = classification_report(y_test, y_pred_dtc)
In [67]:
#随机森林
rd_clf = RandomForestClassifier()
rd_clf.fit(X_train, y_train)

y_pred_rd_clf = rd_clf.predict(X_test)

acc_rd_clf = accuracy_score(y_test, y_pred_rd_clf)
conf = confusion_matrix(y_test, y_pred_rd_clf)
clf_report = classification_report(y_test, y_pred_rd_clf)
In [68]:
#Adaboost模型
ada = AdaBoostClassifier(base_estimator = dtc)
ada.fit(X_train, y_train)

y_pred_ada = ada.predict(X_test)

acc_ada = accuracy_score(y_test, y_pred_ada)
conf = confusion_matrix(y_test, y_pred_ada)
clf_report = classification_report(y_test, y_pred_ada)
In [69]:
#梯度提升树-Gradient Boosting Classifier
gb = GradientBoostingClassifier()
gb.fit(X_train, y_train)

y_pred_gb = gb.predict(X_test)

acc_gb = accuracy_score(y_test, y_pred_gb)
conf = confusion_matrix(y_test, y_pred_gb)
clf_report = classification_report(y_test, y_pred_gb)
In [70]:
#XGBoost模型
xgb = XGBClassifier(booster='gbtree', 
                    learning_rate=0.1, 
                    max_depth=5, 
                    n_estimators=180)

xgb.fit(X_train, y_train)

y_pred_xgb = xgb.predict(X_test)

acc_xgb = accuracy_score(y_test, y_pred_xgb)
conf = confusion_matrix(y_test, y_pred_xgb)
clf_report = classification_report(y_test, y_pred_xgb)
In [71]:
#CatBoost 模型
cat = CatBoostClassifier(iterations=100)
cat.fit(X_train, y_train)

y_pred_cat = cat.predict(X_test)

acc_cat = accuracy_score(y_test, y_pred_cat)
conf = confusion_matrix(y_test, y_pred_cat)
clf_report = classification_report(y_test, y_pred_cat)

Learning rate set to 0.5
0:	learn: 0.4751764	total: 156ms	remaining: 15.5s
1:	learn: 0.4103107	total: 168ms	remaining: 8.23s
2:	learn: 0.3533169	total: 183ms	remaining: 5.91s
3:	learn: 0.3290297	total: 191ms	remaining: 4.59s
4:	learn: 0.2719981	total: 200ms	remaining: 3.8s
5:	learn: 0.2492807	total: 209ms	remaining: 3.28s
6:	learn: 0.2152683	total: 217ms	remaining: 2.88s
7:	learn: 0.1919799	total: 224ms	remaining: 2.58s
8:	learn: 0.1776196	total: 231ms	remaining: 2.34s
9:	learn: 0.1654125	total: 238ms	remaining: 2.14s
10:	learn: 0.1555014	total: 244ms	remaining: 1.97s
11:	learn: 0.1452100	total: 250ms	remaining: 1.83s
12:	learn: 0.1381844	total: 256ms	remaining: 1.71s
13:	learn: 0.1327335	total: 262ms	remaining: 1.61s
14:	learn: 0.1276286	total: 269ms	remaining: 1.52s
15:	learn: 0.1220964	total: 274ms	remaining: 1.44s
16:	learn: 0.1148292	total: 281ms	remaining: 1.37s
17:	learn: 0.1117525	total: 286ms	remaining: 1.3s
18:	learn: 0.1047834	total: 292ms	remaining: 1.25s
19:	learn: 0.1006828	total: 298ms	remaining: 1.19s
20:	learn: 0.0957426	total: 304ms	remaining: 1.14s
21:	learn: 0.0913502	total: 309ms	remaining: 1.1s
22:	learn: 0.0889792	total: 315ms	remaining: 1.05s
23:	learn: 0.0857475	total: 321ms	remaining: 1.02s
24:	learn: 0.0834682	total: 327ms	remaining: 981ms
25:	learn: 0.0827582	total: 333ms	remaining: 947ms
26:	learn: 0.0810976	total: 339ms	remaining: 915ms
27:	learn: 0.0759575	total: 345ms	remaining: 887ms
28:	learn: 0.0713119	total: 351ms	remaining: 859ms
29:	learn: 0.0694644	total: 356ms	remaining: 832ms
30:	learn: 0.0674816	total: 362ms	remaining: 806ms
31:	learn: 0.0653804	total: 368ms	remaining: 781ms
32:	learn: 0.0639664	total: 373ms	remaining: 758ms
33:	learn: 0.0619518	total: 379ms	remaining: 736ms
34:	learn: 0.0594331	total: 385ms	remaining: 716ms
35:	learn: 0.0573412	total: 392ms	remaining: 696ms
36:	learn: 0.0557890	total: 397ms	remaining: 676ms
37:	learn: 0.0534229	total: 403ms	remaining: 658ms
38:	learn: 0.0516835	total: 409ms	remaining: 640ms
39:	learn: 0.0503552	total: 415ms	remaining: 623ms
40:	learn: 0.0488195	total: 421ms	remaining: 606ms
41:	learn: 0.0475821	total: 427ms	remaining: 590ms
42:	learn: 0.0464093	total: 433ms	remaining: 574ms
43:	learn: 0.0452059	total: 439ms	remaining: 559ms
44:	learn: 0.0446879	total: 445ms	remaining: 544ms
45:	learn: 0.0444192	total: 451ms	remaining: 529ms
46:	learn: 0.0434629	total: 457ms	remaining: 515ms
47:	learn: 0.0418914	total: 463ms	remaining: 502ms
48:	learn: 0.0411989	total: 469ms	remaining: 488ms
49:	learn: 0.0409701	total: 474ms	remaining: 474ms
50:	learn: 0.0407533	total: 480ms	remaining: 461ms
51:	learn: 0.0396557	total: 485ms	remaining: 448ms
52:	learn: 0.0392621	total: 491ms	remaining: 436ms
53:	learn: 0.0376227	total: 498ms	remaining: 424ms
54:	learn: 0.0366424	total: 503ms	remaining: 412ms
55:	learn: 0.0347814	total: 509ms	remaining: 400ms
56:	learn: 0.0339543	total: 515ms	remaining: 389ms
57:	learn: 0.0328124	total: 521ms	remaining: 377ms
58:	learn: 0.0320283	total: 527ms	remaining: 366ms
59:	learn: 0.0312842	total: 533ms	remaining: 355ms
60:	learn: 0.0302629	total: 539ms	remaining: 344ms
61:	learn: 0.0300912	total: 544ms	remaining: 333ms
62:	learn: 0.0290183	total: 550ms	remaining: 323ms
63:	learn: 0.0282873	total: 556ms	remaining: 313ms
64:	learn: 0.0279599	total: 561ms	remaining: 302ms
65:	learn: 0.0270445	total: 567ms	remaining: 292ms
66:	learn: 0.0261268	total: 573ms	remaining: 282ms
67:	learn: 0.0258262	total: 579ms	remaining: 273ms
68:	learn: 0.0256831	total: 585ms	remaining: 263ms
69:	learn: 0.0251297	total: 590ms	remaining: 253ms
70:	learn: 0.0249918	total: 596ms	remaining: 244ms
71:	learn: 0.0246868	total: 602ms	remaining: 234ms
72:	learn: 0.0243024	total: 607ms	remaining: 225ms
73:	learn: 0.0235213	total: 613ms	remaining: 215ms
74:	learn: 0.0226188	total: 619ms	remaining: 206ms
75:	learn: 0.0225987	total: 624ms	remaining: 197ms
76:	learn: 0.0223666	total: 630ms	remaining: 188ms
77:	learn: 0.0219795	total: 636ms	remaining: 179ms
78:	learn: 0.0213575	total: 643ms	remaining: 171ms
79:	learn: 0.0210617	total: 649ms	remaining: 162ms
80:	learn: 0.0203673	total: 655ms	remaining: 154ms
81:	learn: 0.0201857	total: 660ms	remaining: 145ms
82:	learn: 0.0194661	total: 666ms	remaining: 136ms
83:	learn: 0.0189876	total: 672ms	remaining: 128ms
84:	learn: 0.0185508	total: 678ms	remaining: 120ms
85:	learn: 0.0179754	total: 684ms	remaining: 111ms
86:	learn: 0.0172582	total: 690ms	remaining: 103ms
87:	learn: 0.0171548	total: 695ms	remaining: 94.8ms
88:	learn: 0.0165666	total: 701ms	remaining: 86.7ms
89:	learn: 0.0158878	total: 707ms	remaining: 78.6ms
90:	learn: 0.0157327	total: 713ms	remaining: 70.5ms
91:	learn: 0.0152266	total: 719ms	remaining: 62.5ms
92:	learn: 0.0150627	total: 724ms	remaining: 54.5ms
93:	learn: 0.0145273	total: 731ms	remaining: 46.6ms
94:	learn: 0.0143670	total: 736ms	remaining: 38.8ms
95:	learn: 0.0138140	total: 742ms	remaining: 30.9ms
96:	learn: 0.0137641	total: 748ms	remaining: 23.1ms
97:	learn: 0.0135317	total: 754ms	remaining: 15.4ms
98:	learn: 0.0132472	total: 759ms	remaining: 7.67ms
99:	learn: 0.0131308	total: 765ms	remaining: 0us
In [72]:
#极端树-Extra Trees Classifier
etc = ExtraTreesClassifier()
etc.fit(X_train, y_train)

y_pred_etc = etc.predict(X_test)

acc_etc = accuracy_score(y_test, y_pred_etc)
conf = confusion_matrix(y_test, y_pred_etc)
clf_report = classification_report(y_test, y_pred_etc)
In [73]:
#LGBM
lgbm = LGBMClassifier(learning_rate = 1)
lgbm.fit(X_train, y_train)

y_pred_lgbm = lgbm.predict(X_test)

acc_lgbm = accuracy_score(y_test, y_pred_lgbm)
conf = confusion_matrix(y_test, y_pred_lgbm)
clf_report = classification_report(y_test, y_pred_lgbm)
In [74]:
#模型11：投票分类器-Voting Classifier
#这个是重点建模：多分类器的投票表决

classifiers = [('Gradient Boosting Classifier', gb), 
               ('Cat Boost Classifier', cat), 
               ('XGboost', xgb),  
               ('Decision Tree', dtc),
               ('Extra Tree', etc), 
               ('Light Gradient', lgbm), 
               ('Random Forest', rd_clf), 
               ('Ada Boost', ada), 
               ('Logistic', lr),
               ('Knn', knn)]

vc = VotingClassifier(estimators = classifiers)
vc.fit(X_train, y_train)

y_pred_vc = vc.predict(X_test)

acc_vtc = accuracy_score(y_test, y_pred_vc)
conf = confusion_matrix(y_test, y_pred_vc)
clf_report = classification_report(y_test, y_pred_vc)

Learning rate set to 0.5
0:	learn: 0.4751764	total: 10.5ms	remaining: 1.04s
1:	learn: 0.4103107	total: 19.7ms	remaining: 964ms
2:	learn: 0.3533169	total: 28.4ms	remaining: 917ms
3:	learn: 0.3290297	total: 35.7ms	remaining: 857ms
4:	learn: 0.2719981	total: 43.2ms	remaining: 822ms
5:	learn: 0.2492807	total: 50.3ms	remaining: 788ms
6:	learn: 0.2152683	total: 57.8ms	remaining: 768ms
7:	learn: 0.1919799	total: 64.3ms	remaining: 739ms
8:	learn: 0.1776196	total: 70.6ms	remaining: 714ms
9:	learn: 0.1654125	total: 76.8ms	remaining: 691ms
10:	learn: 0.1555014	total: 82.8ms	remaining: 670ms
11:	learn: 0.1452100	total: 89.3ms	remaining: 655ms
12:	learn: 0.1381844	total: 95.4ms	remaining: 638ms
13:	learn: 0.1327335	total: 102ms	remaining: 625ms
14:	learn: 0.1276286	total: 108ms	remaining: 611ms
15:	learn: 0.1220964	total: 114ms	remaining: 599ms
16:	learn: 0.1148292	total: 120ms	remaining: 588ms
17:	learn: 0.1117525	total: 127ms	remaining: 577ms
18:	learn: 0.1047834	total: 133ms	remaining: 566ms
19:	learn: 0.1006828	total: 139ms	remaining: 555ms
20:	learn: 0.0957426	total: 144ms	remaining: 543ms
21:	learn: 0.0913502	total: 150ms	remaining: 532ms
22:	learn: 0.0889792	total: 156ms	remaining: 522ms
23:	learn: 0.0857475	total: 162ms	remaining: 514ms
24:	learn: 0.0834682	total: 168ms	remaining: 504ms
25:	learn: 0.0827582	total: 174ms	remaining: 495ms
26:	learn: 0.0810976	total: 180ms	remaining: 485ms
27:	learn: 0.0759575	total: 185ms	remaining: 477ms
28:	learn: 0.0713119	total: 191ms	remaining: 468ms
29:	learn: 0.0694644	total: 197ms	remaining: 459ms
30:	learn: 0.0674816	total: 203ms	remaining: 453ms
31:	learn: 0.0653804	total: 210ms	remaining: 446ms
32:	learn: 0.0639664	total: 216ms	remaining: 439ms
33:	learn: 0.0619518	total: 222ms	remaining: 432ms
34:	learn: 0.0594331	total: 229ms	remaining: 425ms
35:	learn: 0.0573412	total: 235ms	remaining: 418ms
36:	learn: 0.0557890	total: 241ms	remaining: 410ms
37:	learn: 0.0534229	total: 247ms	remaining: 403ms
38:	learn: 0.0516835	total: 253ms	remaining: 395ms
39:	learn: 0.0503552	total: 258ms	remaining: 387ms
40:	learn: 0.0488195	total: 264ms	remaining: 380ms
41:	learn: 0.0475821	total: 270ms	remaining: 372ms
42:	learn: 0.0464093	total: 275ms	remaining: 365ms
43:	learn: 0.0452059	total: 281ms	remaining: 358ms
44:	learn: 0.0446879	total: 287ms	remaining: 351ms
45:	learn: 0.0444192	total: 293ms	remaining: 344ms
46:	learn: 0.0434629	total: 299ms	remaining: 338ms
47:	learn: 0.0418914	total: 305ms	remaining: 331ms
48:	learn: 0.0411989	total: 311ms	remaining: 324ms
49:	learn: 0.0409701	total: 316ms	remaining: 316ms
50:	learn: 0.0407533	total: 322ms	remaining: 309ms
51:	learn: 0.0396557	total: 328ms	remaining: 302ms
52:	learn: 0.0392621	total: 333ms	remaining: 296ms
53:	learn: 0.0376227	total: 339ms	remaining: 289ms
54:	learn: 0.0366424	total: 345ms	remaining: 282ms
55:	learn: 0.0347814	total: 351ms	remaining: 276ms
56:	learn: 0.0339543	total: 357ms	remaining: 269ms
57:	learn: 0.0328124	total: 363ms	remaining: 263ms
58:	learn: 0.0320283	total: 369ms	remaining: 257ms
59:	learn: 0.0312842	total: 375ms	remaining: 250ms
60:	learn: 0.0302629	total: 381ms	remaining: 243ms
61:	learn: 0.0300912	total: 386ms	remaining: 237ms
62:	learn: 0.0290183	total: 392ms	remaining: 230ms
63:	learn: 0.0282873	total: 398ms	remaining: 224ms
64:	learn: 0.0279599	total: 403ms	remaining: 217ms
65:	learn: 0.0270445	total: 409ms	remaining: 211ms
66:	learn: 0.0261268	total: 416ms	remaining: 205ms
67:	learn: 0.0258262	total: 422ms	remaining: 199ms
68:	learn: 0.0256831	total: 429ms	remaining: 193ms
69:	learn: 0.0251297	total: 434ms	remaining: 186ms
70:	learn: 0.0249918	total: 440ms	remaining: 180ms
71:	learn: 0.0246868	total: 446ms	remaining: 173ms
72:	learn: 0.0243024	total: 451ms	remaining: 167ms
73:	learn: 0.0235213	total: 457ms	remaining: 160ms
74:	learn: 0.0226188	total: 463ms	remaining: 154ms
75:	learn: 0.0225987	total: 468ms	remaining: 148ms
76:	learn: 0.0223666	total: 474ms	remaining: 142ms
77:	learn: 0.0219795	total: 480ms	remaining: 135ms
78:	learn: 0.0213575	total: 487ms	remaining: 129ms
79:	learn: 0.0210617	total: 492ms	remaining: 123ms
80:	learn: 0.0203673	total: 498ms	remaining: 117ms
81:	learn: 0.0201857	total: 504ms	remaining: 111ms
82:	learn: 0.0194661	total: 510ms	remaining: 104ms
83:	learn: 0.0189876	total: 516ms	remaining: 98.2ms
84:	learn: 0.0185508	total: 521ms	remaining: 92ms
85:	learn: 0.0179754	total: 527ms	remaining: 85.8ms
86:	learn: 0.0172582	total: 533ms	remaining: 79.7ms
87:	learn: 0.0171548	total: 539ms	remaining: 73.5ms
88:	learn: 0.0165666	total: 545ms	remaining: 67.3ms
89:	learn: 0.0158878	total: 551ms	remaining: 61.2ms
90:	learn: 0.0157327	total: 556ms	remaining: 55ms
91:	learn: 0.0152266	total: 562ms	remaining: 48.9ms
92:	learn: 0.0150627	total: 568ms	remaining: 42.7ms
93:	learn: 0.0145273	total: 574ms	remaining: 36.6ms
94:	learn: 0.0143670	total: 580ms	remaining: 30.5ms
95:	learn: 0.0138140	total: 585ms	remaining: 24.4ms
96:	learn: 0.0137641	total: 591ms	remaining: 18.3ms
97:	learn: 0.0135317	total: 597ms	remaining: 12.2ms
98:	learn: 0.0132472	total: 602ms	remaining: 6.08ms
99:	learn: 0.0131308	total: 608ms	remaining: 0us

基于深度学习keras建模¶

数据预处理和切割¶

In [75]:
from tensorflow.keras.utils import to_categorical

X = pd.concat([cat_df, num_df], axis = 1)
# 转成分类型变量数据
y = to_categorical(df['is_canceled'])
In [76]:
# 切割数据

X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.2)
In [77]:
import tensorflow as tf
import keras
from keras.layers import Dense
from keras.models import Sequential
In [78]:
X.shape[1]
Out[78]:
25
In [79]:
model = Sequential()

model.add(Dense(100, activation="relu",input_shape=(X.shape[1], )))
model.add(Dense(100, activation="relu"))
model.add(Dense(2, activation="sigmoid"))

model.compile(optimizer="adam", 
              loss="binary_crossentropy",
              metrics=["accuracy"])

model_history = model.fit(X_train, 
                          y_train, 
                          validation_data = (X_test, y_test),
                          epochs = 50)

Epoch 1/50
2969/2969 [==============================] - 2s 657us/step - loss: 0.3324 - accuracy: 0.8631 - val_loss: 0.2052 - val_accuracy: 0.9319
Epoch 2/50
2969/2969 [==============================] - 2s 626us/step - loss: 0.1561 - accuracy: 0.9474 - val_loss: 0.1230 - val_accuracy: 0.9582
Epoch 3/50
2969/2969 [==============================] - 2s 610us/step - loss: 0.1179 - accuracy: 0.9615 - val_loss: 0.1209 - val_accuracy: 0.9629
Epoch 4/50
2969/2969 [==============================] - 2s 609us/step - loss: 0.0927 - accuracy: 0.9711 - val_loss: 0.0882 - val_accuracy: 0.9696
Epoch 5/50
2969/2969 [==============================] - 2s 625us/step - loss: 0.0804 - accuracy: 0.9752 - val_loss: 0.0815 - val_accuracy: 0.9717
Epoch 6/50
2969/2969 [==============================] - 2s 637us/step - loss: 0.0732 - accuracy: 0.9769 - val_loss: 0.0654 - val_accuracy: 0.9786
Epoch 7/50
2969/2969 [==============================] - 2s 612us/step - loss: 0.0645 - accuracy: 0.9800 - val_loss: 0.0523 - val_accuracy: 0.9839
Epoch 8/50
2969/2969 [==============================] - 2s 610us/step - loss: 0.0623 - accuracy: 0.9808 - val_loss: 0.0724 - val_accuracy: 0.9762
Epoch 9/50
2969/2969 [==============================] - 2s 607us/step - loss: 0.0585 - accuracy: 0.9818 - val_loss: 0.0522 - val_accuracy: 0.9843
Epoch 10/50
2969/2969 [==============================] - 2s 610us/step - loss: 0.0538 - accuracy: 0.9838 - val_loss: 0.0535 - val_accuracy: 0.9823
Epoch 11/50
2969/2969 [==============================] - 2s 608us/step - loss: 0.0531 - accuracy: 0.9838 - val_loss: 0.0653 - val_accuracy: 0.9793
Epoch 12/50
2969/2969 [==============================] - 2s 610us/step - loss: 0.0504 - accuracy: 0.9846 - val_loss: 0.0643 - val_accuracy: 0.9803
Epoch 13/50
2969/2969 [==============================] - 2s 610us/step - loss: 0.0482 - accuracy: 0.9854 - val_loss: 0.0516 - val_accuracy: 0.9821
Epoch 14/50
2969/2969 [==============================] - 2s 636us/step - loss: 0.0472 - accuracy: 0.9852 - val_loss: 0.0542 - val_accuracy: 0.9831
Epoch 15/50
2969/2969 [==============================] - 2s 609us/step - loss: 0.0436 - accuracy: 0.9863 - val_loss: 0.0522 - val_accuracy: 0.9847
Epoch 16/50
2969/2969 [==============================] - 2s 621us/step - loss: 0.0434 - accuracy: 0.9865 - val_loss: 0.0341 - val_accuracy: 0.9896
Epoch 17/50
2969/2969 [==============================] - 2s 617us/step - loss: 0.0419 - accuracy: 0.9870 - val_loss: 0.0392 - val_accuracy: 0.9881
Epoch 18/50
2969/2969 [==============================] - 2s 613us/step - loss: 0.0423 - accuracy: 0.9872 - val_loss: 0.0383 - val_accuracy: 0.9872
Epoch 19/50
2969/2969 [==============================] - 2s 610us/step - loss: 0.0398 - accuracy: 0.9877 - val_loss: 0.0639 - val_accuracy: 0.9825
Epoch 20/50
2969/2969 [==============================] - 2s 612us/step - loss: 0.0385 - accuracy: 0.9884 - val_loss: 0.0490 - val_accuracy: 0.9824
Epoch 21/50
2969/2969 [==============================] - 2s 626us/step - loss: 0.0381 - accuracy: 0.9881 - val_loss: 0.0498 - val_accuracy: 0.9870
Epoch 22/50
2969/2969 [==============================] - 2s 603us/step - loss: 0.0391 - accuracy: 0.9877 - val_loss: 0.0330 - val_accuracy: 0.9902
Epoch 23/50
2969/2969 [==============================] - 2s 627us/step - loss: 0.0360 - accuracy: 0.9888 - val_loss: 0.0284 - val_accuracy: 0.9903
Epoch 24/50
2969/2969 [==============================] - 2s 596us/step - loss: 0.0373 - accuracy: 0.9887 - val_loss: 0.0367 - val_accuracy: 0.9883
Epoch 25/50
2969/2969 [==============================] - 2s 593us/step - loss: 0.0365 - accuracy: 0.9883 - val_loss: 0.0420 - val_accuracy: 0.9869
Epoch 26/50
2969/2969 [==============================] - 2s 609us/step - loss: 0.0349 - accuracy: 0.9891 - val_loss: 0.0382 - val_accuracy: 0.9880
Epoch 27/50
2969/2969 [==============================] - 2s 612us/step - loss: 0.0323 - accuracy: 0.9902 - val_loss: 0.0292 - val_accuracy: 0.9904
Epoch 28/50
2969/2969 [==============================] - 2s 613us/step - loss: 0.0346 - accuracy: 0.9893 - val_loss: 0.0289 - val_accuracy: 0.9906
Epoch 29/50
2969/2969 [==============================] - 2s 612us/step - loss: 0.0332 - accuracy: 0.9899 - val_loss: 0.0348 - val_accuracy: 0.9886
Epoch 30/50
2969/2969 [==============================] - 2s 613us/step - loss: 0.0322 - accuracy: 0.9900 - val_loss: 0.0560 - val_accuracy: 0.9839
Epoch 31/50
2969/2969 [==============================] - 2s 608us/step - loss: 0.0299 - accuracy: 0.9908 - val_loss: 0.0490 - val_accuracy: 0.9860
Epoch 32/50
2969/2969 [==============================] - 2s 630us/step - loss: 0.0329 - accuracy: 0.9898 - val_loss: 0.0300 - val_accuracy: 0.9903
Epoch 33/50
2969/2969 [==============================] - 2s 600us/step - loss: 0.0299 - accuracy: 0.9908 - val_loss: 0.0283 - val_accuracy: 0.9910
Epoch 34/50
2969/2969 [==============================] - 2s 596us/step - loss: 0.0300 - accuracy: 0.9901 - val_loss: 0.0578 - val_accuracy: 0.9850
Epoch 35/50
2969/2969 [==============================] - 2s 600us/step - loss: 0.0310 - accuracy: 0.9906 - val_loss: 0.0261 - val_accuracy: 0.9919
Epoch 36/50
2969/2969 [==============================] - 2s 600us/step - loss: 0.0293 - accuracy: 0.9907 - val_loss: 0.0364 - val_accuracy: 0.9884
Epoch 37/50
2969/2969 [==============================] - 2s 613us/step - loss: 0.0284 - accuracy: 0.9909 - val_loss: 0.0416 - val_accuracy: 0.9863
Epoch 38/50
2969/2969 [==============================] - 2s 637us/step - loss: 0.0305 - accuracy: 0.9903 - val_loss: 0.0456 - val_accuracy: 0.9877
Epoch 39/50
2969/2969 [==============================] - 2s 632us/step - loss: 0.0281 - accuracy: 0.9915 - val_loss: 0.0235 - val_accuracy: 0.9923
Epoch 40/50
2969/2969 [==============================] - 2s 610us/step - loss: 0.0305 - accuracy: 0.9906 - val_loss: 0.0330 - val_accuracy: 0.9915
Epoch 41/50
2969/2969 [==============================] - 2s 632us/step - loss: 0.0258 - accuracy: 0.9920 - val_loss: 0.0282 - val_accuracy: 0.9906
Epoch 42/50
2969/2969 [==============================] - 2s 604us/step - loss: 0.0270 - accuracy: 0.9913 - val_loss: 0.0486 - val_accuracy: 0.9865
Epoch 43/50
2969/2969 [==============================] - 2s 603us/step - loss: 0.0284 - accuracy: 0.9911 - val_loss: 0.0686 - val_accuracy: 0.9843
Epoch 44/50
2969/2969 [==============================] - 2s 609us/step - loss: 0.0265 - accuracy: 0.9917 - val_loss: 0.0237 - val_accuracy: 0.9936
Epoch 45/50
2969/2969 [==============================] - 2s 603us/step - loss: 0.0269 - accuracy: 0.9916 - val_loss: 0.0621 - val_accuracy: 0.9843
Epoch 46/50
2969/2969 [==============================] - 2s 599us/step - loss: 0.0271 - accuracy: 0.9913 - val_loss: 0.0389 - val_accuracy: 0.9884
Epoch 47/50
2969/2969 [==============================] - 2s 599us/step - loss: 0.0250 - accuracy: 0.9920 - val_loss: 0.0251 - val_accuracy: 0.9924
Epoch 48/50
2969/2969 [==============================] - 2s 599us/step - loss: 0.0258 - accuracy: 0.9919 - val_loss: 0.0336 - val_accuracy: 0.9899
Epoch 49/50
2969/2969 [==============================] - 2s 657us/step - loss: 0.0253 - accuracy: 0.9920 - val_loss: 0.0433 - val_accuracy: 0.9883
Epoch 50/50
2969/2969 [==============================] - 2s 672us/step - loss: 0.0253 - accuracy: 0.9921 - val_loss: 0.0233 - val_accuracy: 0.9928
In [80]:
#指标可视化-loss
train_loss = model_history.history["loss"]
val_loss = model_history.history["val_loss"]

epoch = range(1,51)

loss = pd.DataFrame({"train_loss": train_loss,
                     "val_loss":val_loss
                    })
loss.head()
Out[80]:
train_loss val_loss
0 0.332411 0.205177
1 0.156089 0.123037
2 0.117903 0.120891
3 0.092652 0.088245
4 0.080399 0.081451
In [81]:
train_loss = model_history.history["loss"]
val_loss = model_history.history["val_loss"]

epoch = range(1,51)

loss = pd.DataFrame({"train_loss": train_loss,
                     "val_loss":val_loss
                    })
loss.head()
Out[81]:
train_loss val_loss
0 0.332411 0.205177
1 0.156089 0.123037
2 0.117903 0.120891
3 0.092652 0.088245
4 0.080399 0.081451
In [82]:
fig = px.line(loss, 
        x=epoch, 
        y=['val_loss','train_loss'], 
        title='Train and Val Loss')

fig.show()
In [83]:
#指标可视化-acc
train_acc = model_history.history["accuracy"]
val_acc = model_history.history["val_accuracy"]

epoch = range(1,51)

acc = pd.DataFrame({"train_acc": train_acc,
                     "val_acc":val_acc
                    })

px.line(acc, 
        x=epoch, 
        y=['val_acc','train_acc'], 
        title = 'Train and Val Accuracy',
        template = 'plotly_dark')
In [84]:
#最终预测值

acc_ann = model.evaluate(X_test, y_test)[1]
acc_ann

743/743 [==============================] - 0s 439us/step - loss: 0.0233 - accuracy: 0.9928
Out[84]:
0.9927566647529602

模型对比¶

不同模型的结果对比¶

In [85]:
models = pd.DataFrame({
    'Model' : ['Logistic Regression', 'KNN', 
               'Decision Tree Classifier', 
               'Random Forest Classifier',
               'Ada Boost Classifier',
               'Gradient Boosting Classifier', 
               'XgBoost', 'Cat Boost', 
               'Extra Trees Classifier', 
               'LGBM', 'Voting Classifier','ANN'],
    'Score' : [acc_lr, acc_knn, acc_dtc, 
               acc_rd_clf, acc_ada, acc_gb, 
               acc_xgb, acc_cat, acc_etc, 
               acc_lgbm, acc_vtc, acc_ann]
})


models = models.sort_values(by = 'Score', ascending = True, ignore_index=True)

models["Score"] = models["Score"].apply(lambda x: round(x,4))
models
Out[85]:
Model Score
0 Logistic Regression 0.8135
1 KNN 0.8962
2 Gradient Boosting Classifier 0.9183
3 Decision Tree Classifier 0.9513
4 Ada Boost Classifier 0.9513
5 Extra Trees Classifier 0.9538
6 Random Forest Classifier 0.9562
7 Voting Classifier 0.9672
8 LGBM 0.9740
9 XgBoost 0.9833
10 ANN 0.9928
11 Cat Boost 0.9961
In [86]:
fig = px.bar(models,
       x="Score",
       y="Model",
       text="Score",
       color="Score",
       template="plotly_dark",
       title="Models Comparision"
      )

fig.show()

可以看到Cat Boost分类达到了惊人的99.61%¶

In [ ]: